Cyclin D and cdk4 Are Required for Normal Development ... · Blastula Stage in Sea Urchin Embryos...

MOLECULAR AND CELLULAR BIOLOGY, July 2002, p. 4863–4875 Vol. 22, No. 130270-7306/02/$04.00�0 DOI: 10.1128/MCB.22.13.4863–4875.2002Copyright © 2002, American Society for Microbiology. All Rights Reserved.

Cyclin D and cdk4 Are Required for Normal Development beyond theBlastula Stage in Sea Urchin Embryos

Jennifer C. Moore,1,2 Jan L. Sumerel,1,2 Bradley J. Schnackenberg,1 Jason A. Nichols,1,3

Athula Wikramanayake,4 Gary M. Wessel,5 and William F. Marzluff1,2,3*Program in Molecular Biology and Biotechnology,1 Department of Biochemistry and Biophysics,2 and Department of Biology,3

University of North Carolina, Chapel Hill, North Carolina 27599; Department of Zoology, University of Hawaii, Honolulu, Hawaii968224; and Department of Molecular Cell Biology and Biochemistry, Brown University, Providence, Rhode Island 029125

Received 9 October 2001/Returned for modification 13 November 2001/Accepted 28 March 2002

cdk4 mRNA and protein are constitutively expressed in sea urchin eggs and throughout embryonic devel-opment. In contrast, cyclin D mRNA is barely detectable in eggs and early embryos, when the cell cycles consistof alternating S and M phases. Cyclin D mRNA increases dramatically in embryos at the early blastula stageand remains at a constant level throughout embryogenesis. An increase in cdk4 kinase activity occurs con-comitantly with the increase in cyclin D mRNA. Ectopic expression of cyclin D mRNA in eggs arrestsdevelopment before the 16-cell stage and causes eventual embryonic death, suggesting that activation of cyclinD/cdk4 in cleavage cell cycles is lethal to the embryo. In contrast, blocking cyclin D or cdk4 expression withmorpholino antisense oligonucleotides results in normal development of early gastrula-stage embryos butabnormal, asymmetric larvae. These results suggest that in sea urchins, cyclin D and cdk4 are required fornormal development and perhaps the patterning of the developing embryo, but may not be directly involved inregulating entry into the cell cycle.

Entry of somatic cells into the cell cycle is stringently con-trolled in response to both growth-stimulating and growth-inhibitory signals. One result of this regulation is to ensure thatcells do not commit to replicating their DNA and dividingunless the cell has enough nutrients to complete the process. Inmammalian cells, cyclin-dependent protein kinases (cdk’s) arethe major proteins regulating the cell cycle (42). In order forcdk’s to become active, they must bind their cyclin partner andbe phosphorylated by a cdk-activating kinase (42). Each cyclincontains a region called the cyclin box, which is involved in thebinding of specific cdk’s (30). The best-defined points of thecell cycle that are controlled by cyclins and cdk’s are the G1/Sand G2/M transitions.

Mammalian cells have two G1 cyclins: cyclin D, which bindscdk4 and -6, and cyclin E, which binds cdk2. Cyclin D plays amajor role in the transition of a cell from a resting to a growingstate. Resting mammalian cells have very small amounts ofcyclin D/cdk4 kinase activity, because of both very smallamounts of cyclin D protein and the presence of specific p16/ink4 cdk inhibitors, which bind to free cdk4 and -6 and inhibittheir kinase activity (48, 60). When somatic mammalian cellsare stimulated to reenter the cell cycle, cyclin D mRNA levelsincrease rapidly and remain elevated as long as mitogens arepresent (36). This expression results in the formation of anactive kinase complex composed of cyclin D/cdk4. The majorsubstrate for cyclin D/cdk4 is pRb, and phosphorylation of pRbresults in dissociation of the E2F complex from Rb (7, 17, 29,56). The free E2F proteins in turn stimulate transcription ofgenes required for DNA synthesis (20, 56).

In cycling cells, cyclin D is constitutively synthesized butcontinuously turned over as a result of phosphorylation of aconserved threonine by glycogen synthase kinase-3�, followedby targeting to the proteasome (5). cdk4 is maintained at aconstant level through the cell cycle in all mammalian cellsexamined (37, 38, 40). In contrast, cyclin E/cdk2 activity cyclesin mammalian somatic cells and an increase in cyclin E/cdk2activity are necessary to commit a cell to enter S phase (31).Cyclin E/cdk2 also phosphorylates pRb as well as other mole-cules directly involved in chromosome replication (49).

In addition to control of the cell cycle through regulation ofthe level of cyclin proteins, there are also two families of cdkinhibitors that are important in cell cycle regulation (50). Thep21 family members bind to the cyclin/cdk complexes, whilethe p16/ink4 family members bind directly to cdk4 and cdk6.Homologues of p21 family members are present in all meta-zoans and are important in regulation of the cell cycle inembryonic development in Drosophila melanogaster (4, 33) andCaenorhabditis elegans (24), while the p16 family has only beenidentified in mammals thus far.

The role of the G1 cyclins in the early embryo is less wellunderstood. Many organisms, including sea urchins, gothrough a period of rapid cell division after fertilization, inwhich the cell cycles consist of alternating S phases and mito-ses, with no gap phases. These cell cycles are controlled in partby the oscillation of cyclins A and B (16, 22). Two additionalcell cycle events occur during embryogenesis that are unique todevelopment. The first is the precise timing of the introductionof gap phases, and the second is the complex mechanism thatends maternal control of the embryo and initiates zygotic con-trol. Additionally, the embryo must correctly signal the pat-terning and differentiation of cell lineages within the embryo toform a viable larva.

An unusual characteristic of sea urchins is that their female

* Corresponding author. Mailing address: Program in MolecularBiology and Biotechnology, CB# 7100, University of North Carolina,Chapel Hill, NC 27599. Phone: (919) 962-8920. Fax: (919) 966-6821.E-mail: [email protected].

4863

gametes are stored as haploid eggs. The zygote enters S phasedirectly after fertilization rather than having to first completemeiosis. A large amount of cyclin E is complexed with the cdk2present in the unfertilized egg, and cyclin E/cdk2 levels do notchange during the initial cell cycles (52). Thus, unlike somaticcell cycles, early cell cycles do not require degradation of cyclinE for completion. Although some of the mechanisms thatcontrol the switching of the cell cycles from S/M cycles to cycleswith gap phases have been elucidated in D. melanogaster cells(13, 15), very little is known about this mechanism in verte-brates or other invertebrates. Recent data from D. melano-gaster and C. elegans suggest that cyclin d/cdk4 complexes mayhave some alternative functions other than cell cycle control inthese invertebrates. Specifically, these data show that cyclin Dand cdk4 appear to be more involved in the regulation of cellgrowth and size than in cell cycle progression.

Here we report that in sea urchin embryos, cyclin D mRNAexpression is increased only at the end of cleavage, while cdk4mRNA is constitutively expressed. The increase in cyclin DmRNA is accompanied by an increase in cdk4 kinase activity.Premature expression of cyclin D is lethal to the early embryo,while blocking cdk4 or cyclin D expression results in abnormaldevelopment at the gastrula stage as a result of inappropriatepatterning of the embryo rather than an obvious cell cycledefect. These results demonstrate that cyclinD/cdk4 is essentialfor early embryogenesis in the sea urchin, with the primaryregulatory signal being zygotic expression of cyclin D mRNA.

MATERIALS AND METHODS

Culturing sea urchin embryos. Sea urchin gametes (Strongylocentrotus purpu-ratus and Lytechinus variegatus) were obtained by injecting the urchins with 0.55M KCl. Eggs were collected in seawater at 16°C, and sperm was collected andstored dry at 4°C. Eggs were dejellied by passage through 75-�m Nitex (SefarAmerica, Depew, N.Y.) and packed by centrifugation at 2,000 � g for 5 min. Theeggs were resuspended in seawater at a concentration of 0.5 ml of packedembryos per 100 ml of seawater with 0.1% ampicillin (Sigma, St. Louis, Mo.) withconstant aeration. If the embryos were harvested before hatching, 1 mM para-aminobenzoic acid was added to the dejellied eggs before fertilization to inhibitcross-linking of the fertilization envelope. Embryos that were grown past themesenchyme blastula stage were collected at 18 to 24 h and resuspended at aconcentration of less than 2.5 ml of embryos per liter of seawater for continuedgrowth. Fertilization was checked by light microscopy and was greater than 95%.

Embryo extracts were prepared essentially as described previously (52). Em-bryos were washed twice with 0.55 M KCl. The embryo pellet was resuspendedin 8 volumes of 0.22 M sucrose–10 mM Tris (pH 7.6)–1 mM EGTA–1 mMEDTA–1 mM dithiothreitol. The embryos were pelleted again and resuspendedin 3 volumes of lysis buffer (50 mM Tris [pH 7.6], 0.15 M NaCl, 0.5% NP-40, 1mM dithiothreitol, 1 mM NaVO4, 1 mM phenylmethylsulfonyl fluoride). Theembryos were homogenized with a Dounce homogenizer (tight pestle) andchecked by light microscopy to ensure complete cell lysis.

For analysis of proteins by Western blotting, some extracts were prepared byresuspending the 0.55 M KCl-washed embryos in 1 volume of 0.5 M NaCl,followed by addition of 2 volumes of a 10% sodium dodecyl sulfate (SDS)–5 mMEDTA solution. The embryos were then pipetted up and down several times toimmediately lyse the cells and inactivate proteases and stored at �20°C.

Cloning sea urchin cdk4 and cdk2. A fragment of cdk4 cDNA was isolatedfrom DNA prepared from a sea urchin ovary cDNA library by degenerate PCR(Fig. 1A). The primers used were forward (amino acids HRDLKPQN), 5�-CAY(A/C)GNGAYYTNAARCCN(G/C)ARAA and reverse (amino acidsWYRAPE), 5�-YTCNGGNGCNC(G/T)RTACCA. The amplified fragmentswere cloned, and we obtained 75 clones containing 110-nucleotide fragmentsencoding a portion of sea urchin cdk’s. Three of the 75 clones sequenced en-coded 33 amino acids of the sea urchin homologue of CDK4, based on thepresence of the QMALT sequence that is unique to vertebrate CDK4 andCDK6. The fragment encoding a portion of the sea urchin cdk4 was labeled byPCR in the presence of [�-32P]dCTP and used as a probe to screen an S.

purpuratus ovary library (32); 1.2 million phages were screened, and two cloneswere obtained, one of which encoded the full-length cdk4 protein (accession no.AY044637). The coding region of the S. purpuratus cdk4 was amplified and usedto screen an L. variegatus cDNA library prepared from 4-h (morula stage)embryos (12). Several full-length clones encoding L. variegatus cdk4 were ob-tained and sequenced (accession number AY044638).

We cloned S. purpuratus cdk2 from the S. purpuratus ovary library with thecdk2 from Sphaerechinus granularis (41) as a probe.

Cloning sea urchin cyclin D. The open reading frame of S. purpuratus cdk4 wascloned into the yeast two-hybrid vector pGBT8 and used as bait to screen a 20-hsea urchin cDNA library in the vector pGAD10 (a gift from Bob and LynnAngerer). Since yeast strains expressing sea urchin cdk4 did not transform effi-ciently, a mating yeast two-hybrid screen was performed. The bait and the preyplasmids were transformed into two complementary Saccharomyces cerevisiaemating strains, JP69-4A and JP69-4� (a gift from Niranjan Pandey and TomManiatis). The complementary strains were mated in 1% yeast extract–2% pep-tone–2% dextrose and then plated on synthetic dropout medium lacking LWHand containing 5 mM 3-aminotriazole to select for the most stringent interac-tions. Five positive colonies were obtained and sequenced from 107 coloniesscreened. Four of the five colonies represented portions of the same cDNA andencoded the homologue of the vertebrate cyclin D’s (Fig. 1B).

The fragment obtained from the yeast two-hybrid screen was radiolabeled with[�-32P]dCTP by random primer labeling and used to screen the S. purpuratusphage library as described above. From this screen, three positive colonies wereobtained and sequenced (accession number AF318615). These cDNAs over-lapped, and the final sequence of sea urchin cyclin D is shown in Fig. 1B. The 4-hL. variegatus library was also screened to obtain full-length cyclin D from thatspecies.

In vitro translation. The open reading frames for cdk4 and cyclin D werecloned into the pGEM5 vector, and 1 �g of the purified DNA template was usedin a 50-�l coupled transcription and translation kit (TNT kit; Promega, Madison,Wis.). The reaction mixture contained 25 �l of the TNT rabbit reticulocytelysate, 40 �M amino acid mixture without methionine, and 5 �Ci of [35S]methi-onine-cysteine (Translabel; New England Nuclear, Boston, Mass.). The reactionmixture was incubated at 30°C for 2 h, and the resulting protein was resolved onan SDS–15% polyacrylamide gel, fixed in En3Hance (New England Nuclear),dried, and autoradiographed.

Total RNA isolation. Total RNA was extracted from both eggs and embryos byphenol extraction, essentially as previously described (27). Eggs and embryoswere washed in 0.55 M KCl twice and resuspended in 5 to 10 volumes of 0.25 Msucrose–1 mM EDTA and pelleted by centrifugation. After resuspension, theeggs or embryos were added to an equal volume of a solution containing 2% SDSand 10 mM EDTA (pH 8.0). After mixing well, 1/10 volume of 3 M sodiumacetate (pH 5) was added. This solution was mixed with an equal volume of 4:1.5phenol-chloroform solution and extracted at room temperature for 15 min.Additional phenol-chloroform extractions were performed, and the nucleic acidswere precipitated with 3 volumes of ethanol. The precipitated RNA was col-lected by centrifugation, dissolved at a concentration of about 0.5 mg/ml in asolution containing 0.1% SDS and 1 mM EDTA, extracted one more time withphenol-chloroform, and then ethanol precipitated again. For later-stage embryos(when there is a large amount of chromosomal DNA), the initial extraction wasperformed at 55°C for 10 min to remove chromosomal DNA.

Northern blots. To make probes for Northern blots, the 1,100-bp NcoI frag-ment from cdk4 was purified by preparative gel electrophoresis with a Qiaquikgel extraction kit (Qiagen, Valencia, Calif.), and a 1,600-bp PCR fragment of thecyclin D coding region was amplified. Then 100 ng of the DNA was labeled with50 �Ci of [�-32P]dCTP with a random primed DNA labeling kit (BoehringerMannheim, Indianapolis, Ind.) and purified by elution through a ProbeQuantG-50 microcolumn (Amersham Pharmacia Biotech, Piscataway, N.J.). Samplescontaining equal amounts of total RNA (10 �g/lane) were dissolved in 20 mMMOPS (morpholinepropanesulfonic acid, pH 7.0), 8 mM sodium acetate, 1 mMEDTA (pH 8.0), 6% formaldehyde, and 50% formamide, heated to 65°C for 10min, and separated on an agarose gel containing 6% formaldehyde in 20 mMMOPS (pH 7.0)–8 mM sodium acetate–1 mM EDTA (pH 8.0). The RNA wastransferred to a Hybond nylon membrane (Amersham Pharmacia Biotech, Pis-cataway, N.J.), and the membrane was hybridized with the cdk4 or cyclin D probeat 60°C for 4 h in QuickHybe (Stratagene, La Jolla, Calif.), containing 0.1 mg ofsalmon sperm DNA per ml. After hybridization, the filter was washed twice with2� SSC (1� SSC is 0.15 M NaCl plus 0.015 M sodium citrate)–0.1% SDS at 58°Cfor 15 min and once with 0.1� SSC–0.1% SDS at 58°C for 30 min and autora-diographed.

Antibodies. Peptides corresponding to the C-terminal 15 amino acids of S.purpuratus cyclin D (CVDEVEIITMPSGLS), cdk4 (CDSSQSQDVTPTNKR),

4864 MOORE ET AL. MOL. CELL. BIOL.

and cdk2 (CPYFKDVKMVPPPRL) were synthesized at the University of NorthCarolina Protein Chemistry Facility, coupled to keyhole limpet hemocyanin andused to generate polyclonal antibodies in rabbits (Pocono Rabbit Farm andLaboratory, Canadensis, Pa.). The resulting antibodies were purified with theantigenic peptide cross-linked to a Sulfolink column (Pierce Chemical, Rockford,Ill.). Two milliliters of serum was incubated with the column, and after beingwashed extensively with phosphate-buffered saline, the antibodies were elutedwith 0.1 M glycine-HCl (pH 2.5) and immediately neutralized with 1 M Tris-HCl(pH 9.5). The antibodies recognized the proteins expressed as glutathione S-transferase (GST) fusions in Escherichia coli or in baculovirus by Western blot-ting, although the anti-cdk4 was more than 10-fold more sensitive than theanti-cyclin D and precipitated in vitro-synthesized protein labeled with [35S]me-thionine.

Western blotting. The protein concentration of extracts from sea urchin eggsand embryos was determined by Bradford or bicinchoninic acid analysis (PierceChemical). Equal amounts of total protein (typically 20 �g) from all stages wereresolved on an SDS–12.5% polyacrylamide gel. Western blots were performed asdescribed before (28, 52).

Immunoprecipitations. Samples (10 �l) from in vitro translations were clearedin the presence of 10 �l of protein A-agarose at 4°C for 1 h. After clearing,samples were transferred to a new microcentrifuge tube and immunoprecipitatedwith 1 �l of antibody (approximately 200 �g/ml) at 4°C for 2 h. Followingimmunoprecipitation, the antibodies were incubated with protein A-agarose at4�C for 1 h. The antibody complex was collected by centrifugation, washed in

lysis buffer four times, boiled in sample buffer, and resolved on an SDS–12.5%polyacrylamide gel. The gel was fixed in En3Hance (New England Nuclear), andradiolabeled bands were detected by autoradiography.

Protein kinase assays. Immunoprecipitations were done as previously de-scribed with extracts from sea urchin embryos at the indicated times after fer-tilization (52) and antibodies to either cyclin E or cdk4. After isolation of theantibody complexes with 10 �l of protein A-agarose, the beads were washed inlysis buffer four times. After washing, 25 �l of kinase assay buffer (19) was addedto the beads. Then 4 �g of the C terminus of human pRb conjugated to GST wasadded in the presence of 1 �Ci of [-32P]ATP (specific activity, 3,000 Ci/mmol;New England Nuclear). Because the antibody was produced against a C-terminalpeptide, the specificity of the reaction was determined by the addition of anti-genic peptide to the antibody for 15 min at room temperature before carrying outthe immunoprecipitation.

In situ hybridization. A digoxigenin RNA probe was prepared by in vitrotranscription of cyclin D in the presence of digoxigenin-11-UTP (BoehringerMannheim, Indianapolis, Ind.). Sea urchin eggs and embryos were fixed, washed,rehydrated, and treated with proteinase K as described previously (46). The fixedembryos were hybridized, washed, and visualized as described previously (52).

mRNA microinjections. The sea urchin cyclin D and cdk4 open reading frameswere cloned into the pXFRM expression vector (55) behind an SP6 promoterand in vitro transcribed with the Message Machine transcription kit from Am-bion (Austin, Tex.). L. variegatus eggs were collected and dejellied as describedabove and plated in seawater on plates that had been coated with protamine

FIG. 1. Sea urchin cdk4 and cyclin D. The protein sequences of sea urchin S. purpuratus cdk4 (panel A) and cyclin D (panel B) are comparedto the sequences of cyclin D1 from vertebrates (Homo sapiens, Mus musculus, and Xenopus laevis) and cyclin D from D. melanogaster. In panel A,the amino acid sequences used to construct primers for PCR to identify the cdk genes are underlined. The QMALT sequence characteristic of thevertebrate cdk4/cdk6 genes is boxed. In panel B, the asterisk indicates the threonine identified as the phosphorylation site that determines cyclinD half-life (6). The diamonds indicate the Rb binding motif. Dark shading indicates identical residues, and light shading indicates similar residues.

VOL. 22, 2002 SEA URCHIN CYCLIN D 4865

sulfate. The eggs were fertilized and injected with 0.2 to 1.0 pg of RNA within 5min after fertilization (35). Development of the embryos was monitored by lightmicroscopy.

In vitro translation reaction mixtures containing morpholino oligonucleo-tides. Morpholino oligonucleotides complementary to the 5� untranslated region(UTR) (25 nucleotides before the start codon) of cyclin D and cdk4 weresynthesized by Gene-Tools Inc. (Corvallis, Oreg.). The cdk4 morpholino wasATGCCATCCTCTTCTCTGGTCCCTT, and the cyclin D morpholino wasATCGTCCCACAAGTTCGGTTTGCGG. A total of 10 pmol of either thecdk4, cyclin D, or control morpholino was incubated with in vitro-transcribedRNA at 37°C for 15 min. The RNA morpholino mixture was then in vitrotranslated in the presence of [35S]methionine (TNT kit; Promega, Madison,Wis.), and the products were analyzed by gel electrophoresis and detected byautoradiography.

Morpholino oligonucleotide injection. With the cdk4 and cyclin D morpholi-nos described above, L. variegatus embryos were injected less than 5 min afterfertilization with 1 to 2 pl of a morpholino mixture that contained 2.5 � 10�16

mol (1.5 � 108 molecules) of morpholino, 20% glycerol, and 10 pg of dextran-tetramethyl rhodamine (molecular weight, 10,000; Molecular Probes, Eugene,Oreg.). Development was monitored by light microscopy.

Staining nuclei to determine total cell number. Injected embryos were fixed in90% methanol–50 mM EGTA (pH 6.0) at �20°C (21), washed two times for 5min in 2� SSC, and treated with 100 �g of RNase A per ml in 2� SSC for 20 minat 37°C. After three rinses with 2� SSC, the embryos were incubated in 5 �MSyto11 (Molecular Probes, Eugene, Oreg.) at room temperature for 5 min. Theembryos were then washed three times for 10 min each with 2� SSC. Embryoswere then flattened by a coverslip and mounted in Elvanol (44). Images werecollected with a Zeiss LSM 410 laser scanning confocal system in conjunctionwith a Zeiss Axiovert 100 microscope equipped with a �40/1.3NA objective lens.

RESULTS

Identification of cdk4. We amplified a conserved region ofthe cyclin-dependent kinases with degenerate primers to theunderlined amino acids in Fig. 1A and obtained clones thatcontained 110 nucleotides of the sea urchin homologue of thevarious cdk’s. cdk4 was identified based on the fact that theQMALT sequence (boxed in Fig. 1A), unique to vertebratecdk4 and cdk6, was present in this fragment, and we used thefragment to isolate the S. purpuratus cdk4 cDNA. This cDNAencoded a protein with a predicted molecular mass of 42 kDa(372 amino acids) (Fig. 1A). We arbitrarily designated thisclone cdk4, because it is equally similar to mammalian cdk4and cdk6. We do not know if there is a second member of thisclass of cdk’s in sea urchins.

Sea urchin cdk4 is longer than any other known cdk andcontains a glutamic/aspartic acid-rich region near the C termi-nus. We also cloned cdk4 from the sea urchin L. variegatus withthe S. purpuratus cdk4 as a probe. The L. variegatus cdk4 is alsoa 372-amino-acid protein and contains the C-terminal acidicregion (not shown). It is 94% identical with the S. purpuratuscdk4. Hence, the unusual size of cdk4 is likely a general char-acteristic of the sea urchin cdk4, since the extended C-terminalacidic region is not found in any other known cdk. The longestcDNA clone that we obtained also contained a 5� UTR of 604nucleotides. Since the mRNA for cdk4 is at least 8 kb (seebelow), the 3� UTR must be over 5 kb. There was only limitedsimilarity between the 5� UTRs of the cdk4 cDNAs from thetwo sea urchin species.

Sea urchin cdk4 interacts with mammalian cyclin D andp16. A characteristic of cdk4 is that it binds tightly to cyclin Dand the ink4 class of cdk inhibitors. To determine whether theputative sea urchin cdk4 sequence has the properties of cdk4,we tested its ability to interact with human cyclin D and/or thecdk inhibitor p16 in directed yeast two-hybrid assays. Sea ur-

chin cdk4 was subcloned into the pGBT8 yeast two-hybridvector and transformed into the yeast strain HF7c. Yeast cellsexpressing S. purpuratus cdk4 were then transformed with hu-man cyclin D1, human cyclin D2, or human p16 and cloned asfusion proteins to the Gal4 activation domain in pGAD10.Yeast cells containing both plasmids were streaked onto selec-tive medium and transferred to a nitrocellulose membrane totest for the transcription of the lacZ reporter gene. As shownin Fig. 2, sea urchin cdk4 is able to tightly bind human cyclinsD1 (Fig. 2, panel 6) and D2 (Fig. 2, panel 5) and human p16(Fig. 2, panel 4), and thus we conclude that the sequenceencodes a functional cdk4. As shown in Fig. 2 (panel 2), yeastcells expressing cdk4 fused to the Gal4 DNA binding domain(DBD) in the absence of an interacting protein fused to theGal4 activation domain are unable to grow on medium lackinghistidine.

Functional cloning of sea urchin cyclin D. With the seaurchin cdk4 as bait, 10 million clones were screened for pro-teins that interacted with cdk4 with a yeast two-hybrid cDNAlibrary from 20-h S. purpuratus embryos. To increase the effi-ciency of the screen, a cDNA library that had been trans-formed into the JP69-4A yeast strain was used. The JP69-4�yeast strain was transformed with gal4(DBD)-cdk4, and thelibrary was screened by mating the two yeast strains. Fourpositive clones were identified, each of which contained frag-ments of sea urchin cyclin D, based on their homology to cyclin

FIG. 2. Sea urchin cdk4 interacts with human cyclin D and p16,The S. purpuratus cdk4 open reading frame was fused to the Gal4 DNAbinding domain and transformed into yeast cells expressing the Gal4activation domain (sector 2) or the Gal4 activation domain fused tohuman p16 (sector 4), human cyclin D2 (sector 5), or human cyclin D1(sector 6). Sector 1 is the HF7c yeast strain with no vectors, and sector3 is yeast cells expressing the human cdk4 open reading frame fused tothe Gal4 DNA binding domain transformed with human cyclin D1.The yeast cells were grown on plates lacking leucine and tryptophanand then streaked onto plates lacking leucine, tryptophan, and histi-dine. The colonies were transferred to nitrocellulose and then assayedfor �-galactosidase activity.


D from other organisms. Clones encoding the complete cyclinD coding region were subsequently obtained from both S.purpuratus and L. variegatus by screening phage cDNA librar-ies.

Sea urchin cyclin D is a 303-amino-acid protein which ismost similar to the vertebrate proteins in the cyclin box. Figure1B shows the sequence similarity between cyclin D from vari-ous organisms. In addition to the cyclin box, two key regionsare conserved among the cyclin D proteins in sea urchins andvertebrates. These are the phosphorylation site, CTPT, whichis responsible for determining the half-life of cyclin D in mam-malian cells, and the LXCXE motif, which is responsible forbinding to the Rb family of proteins (Fig. 1B, asterisk anddiamond) (6, 7, 29). Since cyclin D mRNA is about 6.0 kb long,as analyzed by Northern blots (see Fig. 4B), there must be along 3� UTR, as there is in human cyclin D1 (59). The largestclone that we obtained had 1,800 nucleotides of 3� UTR and nopolyadenylation signal. The longest 5� UTR that we obtainedwas 90 nucleotides. The cyclin D from the sea urchin L. var-iegatus was 85% identical to the S. purpuratus cyclin D, differ-ing in 45 amino acids from the S. purpuratus cyclin D, most ofwhich were conservative changes.

Sea urchin cyclin D forms a complex with sea urchin cdk4.A synthetic peptide representing the C terminus of cdk4 wasused to raise a rabbit polyclonal antibody. In order to demon-strate that the cdk4 and cyclin D cloned here from sea urchinsare capable of forming a complex, these proteins were synthe-sized together by in vitro translation, and the reaction mixturewas immunoprecipitated with �-cdk4 and �-cyclin D. The cdk4antibody successfully coimmunoprecipitated cdk4 and cyclin D(Fig. 3A, lane 2), and the precipitation of the two proteins was

blocked by preincubation with the antigenic peptide (Fig. 3A,lane 3). The �-cdk4 did not precipitate cyclin D in the absenceof coexpression of cdk4 (Fig. 3C, lane 2). We coexpressed cdk4and sea urchin cyclin E in the reticulocyte lysate and testedwhether cdk4 could bind cyclin E as assayed by coimmunopre-cipitation with �-cdk4. Only background levels of cyclin E wereprecipitated with the �-cdk4 (Fig. 3B, lane 4). As an additionalcontrol, we used �-cyclin E, which did not immunoprecipitateeither cyclin D or cdk4 (Fig. 3A, lane 4), although it doesprecipitate cyclin E efficiently (52).

We also cloned S. purpuratus cdk2 and prepared an antibodyagainst the C-terminal peptide. We coexpressed cyclin E andcdk2 in the reticulocyte lysate (Fig. 3B, lane 5) and also coex-pressed cyclin D and cdk2 (Fig. 3B, lane 7) and tested whether�-cdk2 could precipitate a cyclin/cdk2 complex. The �-cdk2precipitated both cdk2 and cyclin E (Fig. 3B, lane 6), but didnot coprecipitate cyclin D (Fig. 3B, lane 8).

We prepared several antibodies with peptides or recombi-nant cyclin D, but we only obtained low-affinity antibodies,which precipitated only small amounts of cyclin D from in vitrotranslation reaction mixtures and did not detect cyclin D onWestern blots of embryos. The antibody against the C terminusof cyclin D did coimmunoprecipitate cyclin D and cdk4, butwith much lower efficiency than the �-cdk4 antibody (data notshown). Thus, as in mammalian cells, cdk4 formed a complexwith sea urchin cyclin D but not cyclin E, and cdk2 formed acomplex with cyclin E but not cyclin D.

Expression of cdk4 and cyclin D mRNA during develop-ment. We used Northern blots to determine the amounts ofcdk4 and cyclin D mRNAs present during sea urchin develop-ment. Equal amounts of total RNA (as determined by levels of

FIG. 3. Sea urchin cyclin D and cdk4 form a complex in reticulocyte lysates. (A) Synthetic mRNAs encoding S. purpuratus cyclin D and cdk4were mixed and translated in a rabbit reticulocyte lysate containing [35S]methionine. Equal amounts of lysate were analyzed directly (lane 1) orprecipitated with the anti-cdk4 antibody in the absence (lane 2) or presence (lane 3) of competing antigenic peptide or the antibody to the Cterminus of cyclin E (lane 4). (B) mRNAs encoding S. purpuratus cdk4 and cyclin D (lanes 1 and 2), cdk4 and cyclin E (lanes 3 and 4), cyclin Eand cdk2 (lanes 5 and 6), or cyclin D and cdk2 (lanes 7 and 8) were cotranslated in a rabbit reticulocyte lysate system containing [35S]methionine.The resulting proteins were analyzed directly (lanes 1, 3, 5, and 7) or after precipitation with �-cdk4 (lanes 2 and 4) or �-cdk2 (lanes 6 and 8). Thecyclin E and cdk2 were analyzed directly or after precipitation with �-cyclin E (lane 4). (C) mRNAs encoding S. purpuratus cdk2 and cyclin D werecotranslated in the presence of [35S]methionine (lane 1). The resulting proteins were immunoprecipitated with �-cdk4 (lane 2).


rRNA) were resolved by agarose gel electrophoresis and ana-lyzed with the coding region of cdk4 as a probe. cdk4 mRNAis about 8 to 10 kb long and is present at constant low levels inthe unfertilized egg and through the first 8 h of development(Fig. 4A, lanes 1 to 5). The levels of cdk4 mRNA then in-creased slightly between 8 and 12 h (Fig. 4A, lanes 5 to 7) anddecreased to levels similar to those in the first 8 h of develop-ment by 23 h (mesenchyme blastula; Fig. 4A, lane 8).

In contrast to cdk4 mRNA, the levels of cyclin D mRNAchanged dramatically during development. Cyclin D mRNA (6kb) was present at low levels in the unfertilized egg, and thelevels remained low during the first 8 h of developmentthrough the morula stage (Fig. 4B, lanes 1 to 6). There was alarge increase (at least 10-fold) in cyclin D mRNA in earlyblastula embryos, by 12 h of development, and mRNA levelsremained high throughout blastula (Fig. 4B, lanes 7 and 8) andpersisted through the pluteus stage (Fig. 4C).

In addition to the Northern blots that were performed toquantitate levels of cyclin D message, in situ hybridizationswere done to determine the localization of cyclin D mRNA inthe developing embryos. In agreement with the Northern blots,there were not significant levels of cyclin D mRNA in seaurchin eggs compared to those in later-stage embryos (Fig.5A). In agreement with the Northern blot, cyclin D mRNAlevels were first detectable by in situ hybridization at the blas-tula stage (6 h in L. variegatus), and cyclin D mRNA wasexpressed in all cells at this stage. In mesenchyme blastula- andgastrula-stage embryos, cyclin D mRNA was concentrated inthe vegetal plate and in the ectoderm on the oral side of theembryo (Fig. 5B and C). In pluteus embryos, cyclin D mRNAwas concentrated in the gut and ciliary band (Fig. 5D).

cdk4 protein levels during development. To test whether ourcdk4 antibody could detect small amounts of cdk4, we synthe-sized cdk4 in a coupled in vitro transcription and translationreaction. We readily detected the sea urchin cdk4 by Westernblotting as a single band migrating at 41 kDa (Fig. 6A, lane 1),which was not present in the reticulocyte lysate (Fig. 6A, lane2).

The �-cdk4 also detected a single 41-kDa protein in sea

FIG. 4. Expression of cdk4 and cyclin D mRNA during develop-ment. Total RNA was prepared from S. purpuratus eggs and embryosat the indicated times after fertilization and culturing at 15°C. Equalamounts of total RNA were resolved by Northern blotting, transferredto nitrocellulose, and probed with random primed labeled cdk4 cDNA(A) or cyclin D cDNA (B and C). Eggs are shown in lane 1, andembryos at the indicated times of development are shown in lanes 2 to8 of panel B and lanes 2 to 5 of panel C. The 12-h embryos in panel Bwere at early blastula stage.

FIG. 5. Expression of cyclin D mRNA during embryogenesis by insitu hybridization. L. variegatus embryos at different stages were fixedand analyzed for cyclin D mRNA by in situ hybridization. Eggs (left)and early embryos (A) have no detectable signal until blastula stage(embryo at right), when the signal appears uniform throughout theembryo. At mesenchyme blastula (B), the mRNA is concentrated inthe vegetal plate and the future oral ectoderm. This ectodermal re-gionalization persists through gastrulation (C) to the larval stage (D),and cyclin D mRNA remains concentrated in regions of the ciliaryband. Endoderm labeling also increases during gastrulation to maxi-mal levels in plutei. Panel E is a pluteus larva hybridized with a sensestrand probe. The slight purple color in the egg is due to the pigmentin the egg and not to hybridization.


urchin embryo extracts by Western blotting (Fig. 6B and C).Equal amounts of protein from lysates prepared at differentdevelopmental stages were analyzed by Western blotting todetermine the amount of cdk4 during development. Figure 6B

shows a Western blot of cdk4 over the first cell cycle. Theamount of cdk4 did not change during the first 2 h of devel-opment. Equal amounts of protein from extracts made everyhour for the first 12 h of development were also analyzed for

FIG. 6. Expression of cdk4 protein during development. (A) A coupled transcription-translation reaction in a reticulocyte lysate programmedwith a plasmid expressing cdk4 mRNA (lane 1) or a plasmid (lane 2) was analyzed by Western blotting with the anti-cdk4 antibody. IVT, in vitrotranslation. (B) Extracts were prepared from S. purpuratus eggs (lane 1) and embryos at various times of development. In panel B, embryos wereharvested at the indicated time throughout the first two cell cycles. The embryos divided between 110 and 120 min. (C) Embryos were harvestedat the indicated time after fertilization. Equal amounts of total protein were resolved by SDS gel electrophoresis, transferred to nitrocellulose, andanalyzed by Western blotting with the anti-cdk4 antibody.


cdk4 with �-cdk4 (Fig. 6C). The cdk4 protein levels did notchange significantly during the first 9 h of development andshowed a slight increase between 9 and 13 h that correlatedwith the increase in cdk4 mRNA.

cdk4 kinase activity. To determine whether the cdk4 wasassociated with protein kinase activity, we analyzed the abilityof immunoprecipitates generated with the �-cdk4 antibody tophosphorylate a GST-Rb fusion protein with the same ap-proach that we used previously to detect cyclin E-associatedkinase activity (52). When extracts from 10-h embryos wereanalyzed, cdk4 kinase activity could be detected with a portionof human pRb fused to GST as a substrate (Fig. 7A, lane 1).Preincubation of �-cdk4 with its antigenic peptide blockedprecipitation of kinase activity (Fig. 7A, lane 2). No activity wasobserved toward GST or after incubation of the beads in theextract without antibody (data not shown).

We then used this assay to determine the amount of cdk4during the first 13 h of development. As a control, we alsodetermined the amount of cyclin E-associated kinase activity inthe same extracts. Equal amounts of total embryo extract wereprecipitated with either the cdk4 antibody (Fig. 7B, lanes 1 to5) or the cyclin E antibody (Fig. 7B, lanes 6 to 10), and theimmunoprecipitates were analyzed for protein kinase activity.As reported previously (52) the levels of cyclin E-associatedprotein kinase activity were high in sea urchin eggs and duringthe first cell cycle (Fig. 7B, lanes 6 and 7). The levels of cyclinE-associated kinase decreased dramatically by 10 h of devel-opment and remained low through 13 h (Fig. 7B, lanes 9 and10.

The pattern of cdk4 kinase activity was very different. Onlybackground levels of cdk4 kinase activity (the same as observedwhen the extracts were incubated with protein A beads withoutantibody) were observed in eggs (Fig. 7B, lane 1) and during

the first 2 h of development (Fig. 7B, lanes 1 to 3). Cyclin Dkinase activity was detected by 5 h after development, but thelevels increased dramatically by 12 h of development, after thelarge increase in cyclin D mRNA levels (Fig. 4). Cyclin DmRNA was present in eggs and early embryos, since we iso-lated the cyclin D clones from two-hybrid libraries prepared atthat stage, and low levels of mRNA were detected by Northernblotting (Fig. 4). The levels of cdk4-associated kinase activityclearly increased before the large increase in cyclin D mRNAlevels. The kinase assay is very sensitive, so the small changesobserved at 5 h might not have been reflected in the Northernblot. However, it is also possible that in addition to increasedexpression of cyclin D, there is activation of a small amount ofcyclin D/cdk4 present in early embryos or a small increase incyclin D as a result of posttranscriptional regulatory events.Since the cdk4 protein levels are constant during this time, asanalyzed by Western blotting (Fig. 6), it is likely that thisincrease in cdk4 kinase activity is a result of the synthesis ofcyclin D.

Ectopic expression of cyclin D in early embryos is lethal.Since there is abundant cdk4 protein in the unfertilized eggand probably very little cyclin D (based on the low levels ofcyclin D mRNA and cdk4 kinase activity present), we testedthe effect of precocious cyclin D expression by injecting syn-thetic polyadenylated cyclin D mRNA into L. variegatus zy-gotes within 5 min of fertilization. The premature expression ofcyclin D induced in this manner resulted in the death of 95%of more than 200 injected embryos. The injected embryosarrested and died within 4 h of cyclin D expression, usuallyafter only one or two cell divisions. The death of these embryoswas characterized by extensive blebbing of the cells (Fig. 8).Injection of two negative control RNAs, a cyclin D mRNA thathad a stop codon early in the cyclin D open reading frame andcdk4 mRNA, had no effect on development in more than 200injected embryos. Since this same preparation of cdk4 mRNAwas able to rescue the embryos depleted of cdk4 (see below),overexpression of cdk4 has no effect on embryonic develop-ment.

The effect of the cyclin D mRNA was directly related to theamount of cyclin D mRNA injected. Various amounts of cyclinD mRNA, between 0.2 and 1 pg/embryo, resulted in changingthe time of developmental arrest from the 1-to-2-cell stage (1pg) to the 16-cell stage (0.2 pg). In all cases, when more mRNAwas injected, the embryos completed fewer divisions. The ex-pression of cyclin D early in development, prior to the normaltime of expression, was lethal, probably because of the inap-propriate activation of cyclin D-cdk4 kinase activity, resultingin disruption of the cleavage cell cycles.

Antisense morpholino oligonucleotides block embryonic de-velopment. Antisense morpholino oligonucleotides are a re-cently developed and powerful tool to probe the role of geneactivity during embryonic development (1). This technique uti-lizes long-lived and nontoxic oligonucleotides that stericallyprevent translation of the annealed message (25, 53). Thefunction of a gene can then be rescued by injection of a syn-thetic mRNA that is not complementary to the antisense mor-pholino oligonucleotide. We synthesized synthetic mRNAsthat contained 45 nucleotides of the 5� UTR from L. variegatuscdk4 and the cdk4 coding region from S. purpuratus that con-tained the human �-globin 5� UTR. The resulting mRNAs

FIG. 7. Activity of cdk4 kinase during development. (A) Extractswere prepared from S. purpuratus embryos cultured at 18°C 10 h afterfertilization (early blastula). The extracts were immunoprecipitatedwith �-cdk4 in the absence (lane 1) or presence (lane 2) of antigenicpeptide. The precipitates were assayed for kinase activity with GST-pRb as a substrate (52). (B) Extracts were prepared from a single batchof embryos at the indicated times after fertilization and culturing at18°C. Equal amounts of total protein were immunoprecipitated with�-cdk4 (lanes 1 to 5) or �-cyclin E (lanes 6 to 10), and the immuno-precipitates were assayed for kinase activity. The 10-h embryos were atearly blastula (the same stage as the 12-h embryos in Fig. 4).


were incubated with a 25-nucleotide antisense morpholinocomplementary to L. variegatus cdk4 cDNA from positions�18 to �7 or a control nonspecific morpholino oligonucleo-tide and assayed in an in vitro translation reaction (Fig. 9A).Translation of L. variegatus cdk4 mRNA was inhibited by thecdk4 antisense morpholino (Fig. 9A, lane 2), while translationof the S. purpuratus cdk4 mRNA was not affected by the L.variegatus cdk4 morpholino (Fig. 9A, lane 4).

We performed a similar analysis for L. variegatus and S.purpuratus cyclin D mRNA. The translation of L. variegatuscyclin D mRNA samples containing 67 nucleotides of 5� UTRwas also inhibited by the cyclin D morpholino (complementaryfrom �27 to �2) (Fig. 9B, lane 2), but incubation with thecontrol morpholino had no effect on translation of eithermRNA (Fig. 9B, lanes 1 and 3). S. purpuratus cyclin D mRNAthat contained the human �-globin 5� UTR was not affected bythe L. variegatus cyclin D antisense morpholino (Fig. 9B, lane4). We then used these antisense oligonucleotides and syn-thetic mRNAs to determine whether cdk4 or cyclin D is nec-essary for early development in sea urchins.

Inhibition of cyclin D expression blocks development afterthe blastula stage. To determine if cyclin D was necessary fornormal development, the morpholino oligonucleotide comple-mentary to the 5� UTR of the L. variegatus cyclin D cDNA was

injected into fertilized L. variegatus eggs. The injected embryosnever completed normal gastrulation or developed into normallarvae, although many embryonic structures, including pigmentcells and spicules, were formed (Fig. 10). Although spiculeswere formed, they did not develop into properly organizedskeletons. Although embryos injected with cyclin D morpho-lino became mutant larvae, their development appeared nor-mal through the mesenchyme blastula stage (about 8 to 10 h,since L. variegatus develops faster than S. purpuratus). Morethan 95% of the embryos injected with the cyclin D morpho-lino oligonucleotide developed abnormally, and 95% of em-bryos showed no abnormalities when injected with a controloligonucleotide. Thus, the activation of expression of cyclin Dat the proper time is essential for the correct completion ofgastrulation and the subsequent development into normal plu-tei. Injection of antisense morpholino oligonucleotides againstother cell cycle regulators, cyclin E and cyclin A, resulted inarrested development at an early stage (before 32 cells), aspredicted from the expression pattern of these cell cycle reg-ulators (52).

cdk4 expression is also required for normal development.The delay in cyclin D expression until the blastula stage raisesthe question of whether cdk4 has an essential function prior tothe expression of cyclin D. To test this possibility, a morpho-

FIG. 8. Premature expression of cyclin D mRNA is lethal to sea urchin embryos. Fertilized L. variegatus sea urchin eggs were injected withcyclin D (top) or cdk4 (bottom) mRNA and photographed 2.5 and 5 h after fertilization.


lino oligonucleotide complementary to the 5� end of the cdk4mRNA was injected into one-cell L. variegatus embryos. Likeembryos injected with the cyclin D morpholino, embryos in-jected with the cdk4 morpholino failed to develop into normallarvae. Importantly, the development of the embryos was notaffected until after the blastula stage, even though cdk4 proteinis present throughout development. Since cdk4 protein ispresent in the early embryo, any effect of the morpholinoantisense oligonucleotide requires that the maternal store ofcdk4 protein turn over. It is likely that by the blastula stagethere has been turnover of the maternal cdk4 protein. Anyrequirement for cdk4 in the very early stages of developmentwould not have been detected by this assay. The phenotypes ofthe embryos injected with the cyclin D and cdk4 morpholinooligonucleotides were identical; in each there was developmentof pigment cells and spicules, but a failure to properly com-plete gastrulation and develop normal larvae (Fig. 10). Em-

bryos injected with the control morpholino developed nor-mally.

To determine if the phenotypic alteration resulting from theantisense morpholino oligonucleotide was only due to inhibi-tion of cdk4 activity, we coinjected the cdk antisense morpho-lino oligonucleotide together with a synthetic cdk4 mRNA.This mRNA, which contained the �-globin 5� and 3� UTRs,encoded S. purpuratus cdk4 and did not contain the antisenseoligonucleotide target sequence. Fertilized eggs from the samesea urchin were injected with either the antisense cdk4 mor-pholino, the oligonucleotide plus the synthetic mRNA, or thesynthetic cdk4 mRNA alone. The embryos coinjected with theantisense oligonucleotide and synthetic cdk4 mRNA devel-oped normally, while the embryos injected with the oligonu-cleotide were abnormal in the ways detailed above (Fig. 10). Asshown before, the injection of cdk4 mRNA alone had no effecton development.

From the rescue experiments described above, we concludethat the effects of the cdk4 oligonucleotide were due to theirspecific effect on cdk4 protein expression and not to a generaltoxicity or inhibition of other mRNAs. Since expression of thecyclin D protein early in development is lethal early in devel-opment, we were not able to rescue the effect of the cyclin Doligonucleotide by injection of cyclin D mRNA into fertilizedeggs. The results of the microinjection experiments are sum-marized in Table 1.

Total cell number is not affected by inhibition of expressionof cyclin D or cdk4. To determine if the developmental defectsin the embryos injected with the cdk4 or cyclin D morpholinowere accompanied by a change in the total number of cells, thenuclei of injected embryos were stained with Syto11, and theembryos were flattened with a coverslip, visualized by confocalmicroscopy, and counted. Embryos injected with a controlmorpholino, the cdk4 morpholino, or the cyclin D morpholinoand collected at 24 h postfertilization all contained about 1,400� 200 nuclei (data not shown), showing that there were notlarge differences in cell numbers in the various embryos. Thesedata suggest that inhibiting cyclin D and cdk4 expression didnot result in large differences in the number of cell divisions,suggesting instead that it may play a direct role in the correctpatterning of the embryo.

DISCUSSION

Although cyclins E, A, and B have been shown to playconserved roles in the cell cycle regulation of all metazoans,the role of cyclin D/cdk4 in nonmammalian species is less clear.In both C. elegans and D. melanogaster, genetic analysis indi-cates a role for the cyclin D/cdk4/pRb pathway that does notinvolve the response to mitogens and the passage of a cell fromG1 or G0 to S phase. Instead, cyclin D/cdk4 plays a role inregulating cell size and growth in developing embryos. D. mela-nogaster flies that lack cdk4 (there is only a single homologueof cdk4 and -6 in both D. melanogaster and C. elegans) areviable but are smaller than normal flies and have cells that aresmaller than normal, but show no apparent defects in cell cycleregulation (3, 39). The overexpression of cdk4/cyclin D com-plexes leads to larger than normal cells (3, 39). The degree ofthe enlarged-cell phenotype is proportional to the amount ofcdk4/cyclin D expressed (3).

FIG. 9. Morpholino oligonucleotides inhibit in vitro translation ofcdk4 mRNA (A) and cyclin D mRNA (B). (A) Synthetic L. variegatus(L.v., lanes 1 to 3) or S. purpuratus (S.p., lanes 4 and 5) cdk4 mRNAswere translated in a reticulocyte lysate. The L. variegatus cdk4 anti-sense morpholino oligonucleotide was included in lanes 2 and 4, andthe control morpholino was included in lanes 3 and 5. (B) Synthetic L.variegatus (lanes 1 to 3) or S. purpuratus (lanes 4 and 5) cyclin DmRNAs were translated in a reticulocyte lysate. The L. variegatuscyclin D antisense morpholino oligonucleotide was included in lanes 2and 4, and the control morpholino was included in lanes 3 and 5.


A similar situation is found in C. elegans, in which it has beenshown that neither cdk4 nor cyclin D is required for earlydevelopment (45). With RNA interference, Park and Krauseshowed that loss of cyclin D function results in normal hatchingand development until the L2 stage (45). These animals, how-ever, did have a defect in the germ line, preventing gonad

elongation and resulting in the gonads’ being arrested as a ballof cells (45). Loss of cdk4 function in C. elegans (through RNAinterference or a mutation in cdk4 that disrupts the ATP bind-ing domain) showed a phenotype similar to that for the loss ofcyclin D, with the exception that the embryos that had lost cdk4function were able to survive to later stages (45). These resultsare consistent with cyclin D/cdk4 functioning as a complex inC. elegans and suggest that it does not have an obligatory rolein early development.

A major downstream target of cyclin D/cdk4 in mammaliancells is the pRb family of proteins, which play a critical role inregulation of transcription of genes involved in cell cycle reg-ulation (56). However, it is not known whether cyclin D/cdk4acts through pRb in invertebrates. The D. melanogaster pRbhomologue RbF is not necessary for the introduction of G1

phases during embryogenesis, but RbF is necessary for themaintenance of cells in the G1 phase of the cell cycle (8).However, the E2F/DP proteins in D. melanogaster, majordownstream targets of pRb, are clearly involved in cell cycleregulation in D. melanogaster, regulating transcription of manyof the same genes that are E2F regulated in mammalian cellsduring the cell cycle (9–11, 47). C. elegans mutants carrying amutation in the pRb homologue show inappropriate hyperac-tivation of the Ras pathway and do not have an obvious cellcycle phenotype (34). The same Ras pathway phenotype is

FIG. 10. Cyclin D and cdk4 are required for the normal development of sea urchin larvae. Fertilized L. variegatus sea urchin eggs were injectedwith a morpholino antisense control oligonucleotide, a morpholino antisense oligonucleotide to cyclin D, a morpholino antisense oligonucleotideto cdk4, or a mixture of the antisense oligonucleotide and a synthetic S. purpuratus cdk4 mRNA. Embryos were photographed 24 and 48 h afterfertilization. Two examples of the embryos injected with antisense cyclin D or antisense cdk4 are shown.

TABLE 1. Effect of injection of antisense morpholinooligonucleotides and synthetic mRNAs on developmenta

Injected nucleic acid Phenotype (% of embryos)

Control morpholino.................Normal (95 � 5)Cyclin D morpholino...............Abnormal gastrulation (85 � 10)cdk4 morpholino ......................Abnormal gastrulation (85 � 7)cdk4 mRNA..............................Normal (95 � 3)

Cyclin D mRNA ......................Arrest by 4-cell stage (75 � 10); arrest

by 16-cell stage (90 � 10)cdk4 morpholino � cdk4

mRNA ...................................Normal (75 � 14)

a One picoliter of solution containing either 250 �M morpholino oligonucle-otide or 0.5 to 0.7 pg of synthetic polyadenylated and capped mRNA was injectedinto L. variegatus zygotes within 5 min after fertilization. The embryos wereallowed to develop for 48 h, and the development of the embryos was monitoredby light microscopy. Each oligonucleotide or mRNA was injected into at leastfive different sets of zygotes, between 200 and 600 embryos were scored for eachtreatment, and the percentage of embryos showing the indicated phenotype isgiven in parentheses, with the range of individual experiments. The abnormalgastrulation phenotype is that shown in Fig. 8B.


found in C. elegans with mutant E2F and DP, two proteins thatare downstream targets of pRb, suggesting that in C. elegansthis pathway is not primarily involved in cell cycle regulation(2, 43).

We have detailed here the cloning of S. purpuratus cdk4 andits partner cyclin D. Both of these proteins are homologous totheir vertebrate counterparts. In addition to interacting withcyclin D, sea urchin cdk4 interacts with mammalian p16, acdk4-specific inhibitor, even though members of this class ofinhibitor have not been reported in any organisms exceptmammals. We have also shown that cdk4 and cyclin D interactonly with each other and not with other cyclins and cdk’s, suchas cyclin E and cdk2. The time of induction of cyclin D duringsea urchin development corresponds to the time of reductionin cyclin E (52) and may coincide with the change in cell cyclecontrol from cycles that lack gap phases to cycles that includegap phases. Although cdk4 protein is present in eggs andduring embryogenesis, cyclin D mRNA does not accumulatesignificantly until 10 h after fertilization, and cdk4 kinase ac-tivity is initiated only at 5 h postfertilization and increases asthe levels of cyclin D mRNA increase. Thus, cdk4 is probablynot active until the blastula stage. In support of this, knockingout cdk4 with an antisense morpholino oligonucleotide did notaffect development up to the blastula stage, but disrupted laterdevelopment. The same phenotype was observed when expres-sion of cyclin D was prevented, consistent with the possibilitythat the critical function of cdk4 is to form a complex withcyclin D.

In mammalian tissue culture cells, cyclin D has been shownto be the critical factor induced by mitogenic signals that isessential for the cells to reenter the cell cycle and progress toS phase (36, 58). A primary effect of cdk4 is phosphorylation ofpRb, ultimately resulting in the activation of transcription of anumber of genes required for entry into S phase (56). Earlydeveloping embryos likely do not need transcription and trans-lation of the genes involved in entry into S phase, since theseproteins and mRNAs are constitutively present. In addition,since cells progress immediately from mitosis into the next Sphase, active cyclin D/cdk4 is not necessary to stimulate exitfrom G1 phase. Since conventionally cycling cells need to makea decision after mitosis when to commit to enter the next Sphase, cyclin D may play a key role in that decision. The signalsthat stop cells after mitosis likely are not produced in therapidly cycling cells.

There are three cyclin D’s in mammalian cells, in contrast tothe single cyclin D in C. elegans and D. melanogaster, and theseare likely to have overlapping functions. The phenotypes ofcyclin D1 and cyclin D2 knockouts in mice were relatively mild(18, 51). Mice lacking cyclin D1 were smaller than their litter-mates and showed defects in their retinas and mammaryglands, but there were no severe developmental defects (18).Similarly, mice lacking cyclin D2 were indistinguishable fromtheir littermates, but mutant females were infertile due toovarian failure (51). Cyclin D is first expressed in largeamounts just prior to gastrulation in mouse embryos (57),similar to its delayed expression in embryogenesis in both seaurchins and frogs (54). The lack of severe phenotypes andembryonic lethality is likely a result of the ability of othermembers of the cyclin D family to compensate for the loss ofcyclin D1 or cyclin D2.

Ectopic expression of cyclin D in cleavage embryos as aresult of injecting cyclin D mRNA into sea urchin eggs disruptsthe cleavage cell cycles, resulting in embryonic death, with adirect correlation to the amount of mRNA injected. A similarobservation has been made in frog embryos; premature expres-sion of cyclin D blocks cleavage of frog embryos (54). Theeffect of premature expression of cyclin D could be due tostorage of many of the components (except for cyclin D) nec-essary for the switch in cell cycle regulation, from S/M cycles tocycles with gap phases, in the egg and early embryo. This switchmay involve the synthesis of cyclin D as well as changes in othercell cycle regulators, such as a reduction in cyclin E (23, 26, 52).A precedent for this type of regulation has been seen in D.melanogaster embryos, in which cyclins A and B are titratedaway during the continuous S/M cycles, allowing the embryo todevelop an interphase lag and activating degradation of the D.melanogaster cdc25 homologue (14).

The mutant embryos that develop as a result of cyclin D orcdk4 loss of function do not differ significantly from controlembryos in total cell number. This suggests that even in laterstages of sea urchin embryogenesis, cyclin D/cdk4 kinase ac-tivity is not directly involved in entry into DNA synthesis fromG1. Like the role of cdk4/cyclin D in D. melanogaster, in seaurchins cdk4/cyclin D may be more intimately involved in cellgrowth and the patterning of embryonic cells. This is supportedby the demonstration that loss of cyclin D or cdk4 function onlyaffects some cell lineages in the developing embryo. Somelineages found in larvae, such as pigment cells, are clearlypresent.

Thus, while the cyclins A, B, and E have had conservedfunctions throughout metazoan evolution, the biological roleof cyclin D appears to have changed. Unlike C. elegans and D.melanogaster, sea urchins have a clear requirement for cyclinD/cdk4 to complete early embryonic development, but it doesnot appear that this results from a disruption of a function incell cycle regulation. One possibility is that this effect is due torelatively subtle perturbation of the cell cycle or the incorrectsignaling of differentiation in specific lineages, and this will bethe subject of future investigations.

ACKNOWLEDGMENTS

This work was supported by NIH grants GM 58921 (W.F.M.) andHD28152 and HD01170 (G.M.W.) and a grant from the NSF (IBN-9816683) to G.M.W. B.J.S. was supported by National Cancer InstituteTraining Grant T32CA09156 and postdoctoral fellowshipF32GM20151.

We thank Bob and Lynne Angerer for advice on microinjection ofthe antisense oligonucleotides, Randy Thresher and Lilly Reed forassistance with microinjection and photography, and Bob Duronio,Yue Xiong, and Dave McClay for many helpful discussions.

REFERENCES

1. Audic, Y., B. Boyle, M. Slevin, and R. S. Hartley. 2001. Cyclin E morpholinodelays embryogenesis in Xenopus. Genesis 30:107–109.

2. Ceol, C. J., and H. R. Horvitz. 2001. dpl-1 DP and elf-1 E2F act with lin-35Rb to antagonize ras signaling in C. elegans vulval development. Mol. Cell3:461–474.

3. Datar, S. A., H. W. Jacobs, A. F. De la Cruz, C. F. Lehner, and B. A. Edgar.2000. The D. melanogaster cyclin D-Cdk4 complex promotes cellular growth.EMBO J. 19:4543–4554.

4. De Nooij, J. C., M. A. Letendre, and I. K. Hariharan. 1996. A cyclin-dependent kinase inhibitor, dacapo, is necessary for timely exit from the cellcycle during D. melanogaster embryogenesis. Cell 87:1237–1247.

5. Diehl, J. A., M. G. Cheng, M. F. Roussel, and C. J. Sherr. 1998. Glycogen


synthase kinase 3� regulates cyclin D1 proteolysis and subcellular localiza-tion. Genes Dev. 12:3499–3511.

6. Diehl, J. A., F. Zindy, and C. J. Sherr. 1997. Inhibition of cyclin D1 phos-phorylation on threonine-286 prevents its rapid degradation via the ubiquin-tin-proteasome pathway. Genes Dev. 11:957–972.

7. Dowdy, S. F., P. W. Hinds, K. Louie, S. I. Reed, A. Arnold, and R. A.Weinberg. 1993. Physical interaction of the retinoblastoma protein withhuman D cyclins. Cell 73:499–511.

8. Du, W., and N. Dyson. 1999. The role of RBF in the introduction of G1regulation during D. melanogaster embryogenesis. EMBO J. 18:916–925.

9. Duronio, R. J., P. C. Bonnette, and P. H. O’Farrell. 1998. Mutations of theDrosophila dDP, dE2F, and cyclin E genes reveal distinct roles for theE2F-DP transcription factor and cyclin E during the G1-S transition. Mol.Cell. Biol. 18:141–151.

10. Duronio, R. J., and P. H. O’Farrell. 1995. Developmental control of the G1to S transition in D. melanogaster: cyclin E is a limiting downstream target ofE2F. Genes Dev. 9:1456–1468.

11. Duronio, R. J., P. H. O’Farrell, J.-E. Xie, A. Brook, and N. Dyson. 1995. Thetranscription factor E2F is required for S phase during D. melanogasterembryogenesis. Genes Dev. 9:1445–1455.

12. Edelmann, L., L. X. Zheng, Z. F. Wang, W. F. Marzluff, G. M. Wessel, andG. Childs. 1998. The TATA binding protein in the sea urchin embryo ismaternally derived. Dev. Biol. 204:293–304.

13. Edgar, B. A. 1994. Cell cycle. Cell-cycle control in a developmental context.Curr. Biol. 4:522–524.

14. Edgar, B. A., and S. A. Datar. 1996. Zygotic degradation of two maternalCdc25 mRNAs terminates D. melanogaster’s early cell cycle program. GenesDev. 10:1966–1977.

15. Edgar, B. A., and C. F. Lehner. 1996. Developmental control of cell cycleregulators: a fly’s perspective. Science 274:1646–1652.

16. Edgar, B. A., F. Sprenger, R. J. Duronio, P. Leopold, and P. H. O’Farrell.1994. Distinct molecular mechanism regulate cell cycle timing at successivestages of D. melanogaster embryogenesis. Genes Dev. 8:440–452.

17. Ewen, M. E., H. K. Sluss, C. J. Sherr, H. Matsushime, J. Kato, and D. M.Livingston. 1993. Functional interactions of the retinoblastoma protein withmammalian D-type cyclins. Cell 73:487–497.

18. Fantl, V., G. Stamp, A. Andrews, I. Rosewell, and C. Dickson. 1995. Micelacking cyclin D1 are small and show defects in eye and mammary glanddevelopment. Genes Dev. 9:2364–2372.

19. Gutierrez, C., Z.-S. Guo, W. Burhans, M. L. DePamphilis, J. Farrell-Towt,and G. Ju. 1988. Is c-Myc protein directly involved in DNA replication?Science 240:1202–1202.

20. Hamel, P. A., B. L. Gallie, and R. A. Phillips. 1992. The retinoblastomaprotein and cell cycle regulation. Trends Genet. 8:180–185.

21. Harris, P. J. 1986. Cytology and immunocytochemistry. Methods Cell Biol.27:243–262.

22. Hartley, R. S., R. E. Rempel, and J. L. Maller. 1996. In vivo regulation of theearly embryonic cell cycle in Xenopus. Dev. Biol. 173:408–419.

23. Hartley, R. S., J. C. Sible, A. L. Lewellyn, and J. L. Maller. 1997. A role forcyclin E/Cdk2 in the timing of the midblastula transition in Xenopus em-bryos. Dev. Biol. 188:312–321.

24. Hong, Y., R. Roy, and V. Ambros. 1998. Developmental regulation of acyclin-dependent kinase inhibitor controls postembryonic cell cycle progres-sion in Caenorhabditis elegans. Development 125:3585–3597.

25. Howard, E. W., L. A. Newman, D. W. Oleksyn, R. C. Angerer, and L. M.Angerer. 2001. SpKrl: a direct target of �-catenin regulation required forendoderm differentiation in sea urchin embryos. Development 128:365–375.

26. Howe, J. A., and J. W. Newport. 1996. A developmental timer regulatesdegradation of cyclin E1 at the midblastula transition during Xenopus em-bryogenesis. Proc. Natl. Acad. Sci. USA 93:2060–2064.

27. Jarvis, J. W., and W. F. Marzluff. 1989. The early and late sea urchin histoneH4 mRNAs respond differently to inhibitors of DNA synthesis. Dev. Biol.132:325–330.

28. Jenkins, C. W., and Y. Xiong. 1995. Immunoprecipitations and immunoblot-ting in cell cycle studies, p. 250–263. In M. Pagano (ed.), Cell cycle: materialsand methods. Springer-Verlag, New York, N.Y.

29. Kato, J., H. Matsushime, S. W. Hiebert, M. E. Ewen, and C. J. Sherr. 1993.Direct binding of cyclin D to the retinoblastoma gene product (pRb) andpRb phosphorylation by the cyclin D-dependent kinase CDK4. Genes Dev.7:331–342.

30. Kobayashi, H., E. Stewart, R. Poon, J. P. Adamczewski, J. Gannon, and T.Hunt. 1992. Identification of the domains in cyclin A required for binding to,and activation of, p34cdc2 and p32cdk2 protein kinase subunits. Mol. Biol.Cell 3:1279–1294.

31. Kosterin, O. E., V. S. Bogdanova, F. L. Gorel, S. M. Rozov, Y. A. Trusov, andV. A. Berdnikov. 1994. Histone H1 of the garden pea (Pisum sativum L.);composition, developmental changes, allelic polymorphism and inheritance.Plant Sci. 101:189–202.

32. Laidlaw, M., and G. M. Wessel. 1994. Cortical granule biogenesis is activethroughout oogenesis in sea urchins. Development 120:1325–1333.

33. Lane, M. E., K. Sauer, K. Wallace, Y. N. Jan, C. F. Lehner, and H. Vaessin.

1996. Dacapo, a cyclin-dependent kinase inhibitor, stops cell proliferationduring D. melanogaster development. Cell 87:1225–1235.

34. Lu, X. W., and H. R. Horvitz. 1998. lin-35 and lin-53, two genes that antag-onize a C. elegans Ras pathway, encode proteins similar to Rb and its bindingprotein RbAp48. Cell 95:981–991.

35. Mao, C. A., A. H. Wikramanayake, L. Gan, C. K. Chuang, R. G. Summers,and W. H. Klein. 1996. Altering cell fates in sea urchin embryos by overex-pressing SpOtx, an orthodenticle-related protein. Development 122:1489–1498.

36. Matsushime, H., M. F. Roussel, R. A. Ashmun, and C. J. Sherr. 1991.Colony-stimulating factor 1 regulates novel cyclins during the G1 phase ofthe cell cycle. Cell 65:701–713.

37. Matushansky, I., F. Radparvar, and A. I. Skoultchi. 2000. Manipulating theonset of cell cycle withdrawal in differentiated erythroid cells with cyclin-dependent kinases and inhibitors. Blood 96:2755–2764.

38. Matushansky, I., F. Radparvar, and A. I. Skoultchi. 2000. Reprogrammingleukemic cells to terminal differentiation by inhibiting specific cyclin-depen-dent kinases in G1. Proc. Natl. Acad. Sci. USA 97:14317–14322.

39. Meyer, C. A., H. W. Jacobs, S. A. Datar, W. Du, B. A. Edgar, and C. F.Lehner. 2000. D. melanogaster Cdk4 is required for normal growth and isdispensable for cell cycle progression. EMBO J. 19:4533–4542.

40. Meyerson, M., and E. Harlow. 1994. Identification of G1 kinase activity forcdk6, a novel cyclin D partner. Mol. Cell. Biol. 14:2077–2086.

41. Moreau, J. L., F. Marques, A. Barakat, P. Schatt, J. C. Lozano, G. Peaucel-lier, A. Picard, and A. M. Genevière. 1998. Cdk2 activity is dispensable forthe onset of DNA replication during the first mitotic cycles of the sea urchinearly embryo. Dev. Biol. 200:182–197.

42. Morgan, D. O. 1995. Principles of CDK regulation. Nature 374:131–134.43. Page, B. D., S. Guedes, D. Waring, and J. R. Priess. 2001. The C. elegans

E2F- and DP-related proteins are required for embryonic asymmetry andnegatively regulate Ras/MAPK signaling. Mol. Cell 7:451–460.

44. Palazzo, R. E., and J. V. Vogel. 1999. Isolation of centrosomes from Spisulasolidissima oocytes. Methods Cell Biol. 61:35–56.

45. Park, M., and M. W. Krause. 1999. Regulation of postembryonic G1 cellcycle progression in Caenorhabditis elegans by a cyclin D/CDK-like complex.Development 126:4849–4860.

46. Ransick, A., S. Ernst, R. J. Britten, and E. H. Davidson. 1993. Whole mountin situ hybridization shows Endo 16 to be a marker for the vegetal plateterritory in sea urchin embryos. Mech. Dev. 42:117–124.

47. Royzman, I., A. J. Whittaker, and T. L. Orr-Weaver. 1997. Mutations in D.melanogaster DP and E2F distinguish G1-S progression from an associatedtranscriptional program. Genes Dev. 11:1999–2011.

48. Serrano, M., G. J. Hannon, and D. Beach. 1993. A new regulatory motif incell cycle control cawith specific inhibition of cyclin D/CDK4. Nature 366:704–707.

49. Shen, X. T., G. Mizuguchi, A. Hamiche, and C. Wu. 2000. A chromatinremodelling complex involved in transcription and DNA processing. Nature406:541–544.

50. Sherr, C. J., and J. M. Roberts. 1999. CDK inhibitors: positive and negativeregulators of G1-phase progression. Genes Dev. 13:1501–1512.

51. Sicinski, P., J. L. Donaher, Y. Geng, S. B. Parker, H. Gardner, M. Y. Park,R. L. Robker, J. S. Richards, L. K. McGinnis, J. D. Biggers, J. J. Eppig, R. T.Bronson, S. J. Elledge, and R. A. Weinberg. 1996. Cyclin D2 is an FSH-responsive gene involved in gonadal cell proliferation and oncogenesis. Na-ture 384:470–474.

52. Sumerel, J. L., J. C. Moore, B. J. Schnackenberg, J. A. Nichols, J. C.Canman, G. M. Wessel, and W. F. Marzluff. 2001. Cyclin E and its associatedcdk activity do not cycle during early embryogenesis of the sea urchin. Dev.Biol. 234:425–440.

53. Summerton, J., and D. Weller. 1997. Morpholino antisense oligomers: de-sign, preparation, and properties. Antisense Nucleic Acid Drug Dev. 7:187–195.

54. Taieb, F., I. Chartrain, S. Chevalier, O. Haccard, and C. Jessus. 1997. CyclinD2 arrests Xenopus early embryonic cell cycles. Exp. Cell Res. 237:338–346.

55. Wang, Z.-F., T. C. Ingledue, Z. Dominski, R. Sanchez, and W. F. Marzluff.1999. Two Xenopus proteins that bind the 3� end of histone mRNA: impli-cations for translational control of histone synthesis during oogenesis. Mol.Cell. Biol. 19:835–845.

56. Weinberg, R. A. 1995. The retinoblastoma protein and cell cycle control. Cell81:323–330.

57. Wianny, F., F. X. Real, C. L. Mummery, M. Van Rooijen, J. Lahti, J.Samarut, and P. Savatier. 1998. G1-phase regulators, cyclin D1, cyclin D2,and cyclin D3: up-regulation at gastrulation and dynamic expression duringneurulation. Dev. Dyn. 212:49–62.

58. Won, K. A., Y. Xiong, D. Beach, and M. Z. Gilman. 1992. Growth-regulatedexpression of D-type cyclin genes in human diploid fibroblasts. Proc. Natl.Acad. Sci. USA 89:9910–9914.

59. Xiong, Y., T. Connolly, B. Futcher, and D. Beach. 1991. Human D-typecyclin. Cell 65:691–699.

60. Xiong, Y., H. Zhang, and D. Beach. 1993. Subunit rearrangement of thecyclin-dependent kinases is associated with cellular transformation. GenesDev. 7:1572–1583.


Cyclin D and cdk4 Are Required for Normal Development ... · Blastula Stage in Sea Urchin Embryos...

Documents

Transcript of Cyclin D and cdk4 Are Required for Normal Development ... · Blastula Stage in Sea Urchin Embryos...